Rate-Constrained Simulation and Source Coding IID Sources

TL;DR

This paper establishes necessary conditions for optimal source coding and simulation of memoryless sources, proposing a trellis-based design technique that outperforms existing methods in certain scenarios.

Contribution

It introduces a novel design approach for rate-constrained source coding and simulation based on necessary conditions, with experimental validation showing improved performance.

Findings

The proposed method achieves comparable or better performance than existing techniques.

Experimental results demonstrate significant improvements on common examples.

The approach is inspired by classic random coding and alphabet-constrained methods.

Abstract

Necessary conditions for asymptotically optimal sliding-block or stationary codes for source coding and rate-constrained simulation of memoryless sources are presented and used to motivate a design technique for trellis-encoded source coding and rate-constrained simulation. The code structure has intuitive similarities to classic random coding arguments as well as to ``fake process'' methods and alphabet-constrained methods. Experimental evidence shows that the approach provides comparable or superior performance in comparison with previously published methods on common examples, sometimes by significant margins.